112 SELECTION IN CLADOCERA ON THE BASIS OF 



For the following June- July period, during which the minus strain 

 was within 20 seconds of being as reactive as the plus strain, it again 

 was slightly the more vigorous. But it is noteworthy that in the 

 period intervening between the two periods just referred to the 

 minus strain was superior in vigor to the plus strain by a difference 

 3 times as great as in the other two periods, and yet the reaction-time 

 for the minus strain was 101 seconds greater than that for the plus 

 strain. For the last two-month period (December 191 6- January 

 1917), in which the reaction- time mean for the minus approached 

 that for the plus strain, the minus was considerably the more vigorous. 

 It is noteworthy here, however, that in the period just preceding 1 

 the minus strain showed greater superiority in vigor than at any 

 other time, and yet its reaction-time differed from that for the plus 

 by a larger margin (629 seconds) than for any other period of the 

 experiment. This analysis might be prolonged further, but enough 

 has been seen to make clear that there is at most a very slight rela- 

 tion between the reproductive index and the mean reaction-time. 



Of the points discussed above, (3) and doubtfully (4) and (5) 

 indicate a possible effect of vigor upon reaction-time in the plus 

 strain (but not in the minus strain) i. e., the greater the vigor the 

 more reactive the individuals and the lower the reaction-time. But 

 this bit of evidence is not very convincing, since the numerical differ- 

 ences in the comparisons made are small and the differences in 

 averages are not all in the right direction to bear out such a relation- 

 ship. On the other hand, (1), (2), and (8) fail to show the appearance 

 of an effect of mean vigor upon mean reaction-time, while (5) (in 

 part), (6), and (7) seem to indicate the reverse effect, that greater 

 vigor is associated with lessened reactiveness and a higher reaction- 

 time. 



Obviously, the best method of determining the relation, if any, 

 between reproductive vigor and mean reaction-time is by statistical 

 correlation. The mean reaction-times and mean reproductive 

 indices, both by two-month averages, were correlated for the data for 

 the Line 757 strains, using as the terms of the correlation the repro- 

 ductive index of the mother and the mean reaction-time of her first 

 brood of offspring. The correlation is 0.1 109 0.127 for the plus 

 strain and 0.0612 0.129 for the minus strain. In order to have a 

 larger series, the data for both strains of Line 757 were thrown 

 together. 2 The resulting correlation is -0.12845 0.0902. These 

 correlations are low, and, judged by their probable errors, are not of 

 statististical significance. 



J The numbers of individuals tested during this period were abnormally small, so that these 

 averages are not as trustworthy as most of the others, but the major part of this difference can 

 not be regarded aa accidental. 



1 Since the two strains are different in their reactiveness to light, this combination of data is 

 open to criticism, although it is reasonable to suppose that variations in vigor should influence 

 the reaction-time in the two strains similarly. 



